4

Has anyone ever seen machine learning (ML) used to assist a Max Flow algorithm?

I have a very large directed graph that has some fractal characteristics, meaning that this large graph can be roughly split into smaller ones.

I was wondering if I could run a Max Flow into one of this smaller instances and use this knowledge to train an ML system as to what nodes participate in the max flow solution and what nodes don't (feeding info such as node degree, reachability to some other node, etc, etc).

Thus, I could then use this trained system to give me hints of what nodes should I prune from the large graph or to what edges should I push flow... Has anyone ever heard of anything like that?

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
rafbrl
  • 41
  • 1
  • I've never even heard of max flow problems, but I hope you get an answer because it seems interesting. But it sounds like what you really want is a genetic algorithm (try stuff randomly and keep what works), and in my understanding this is the kind of problem where genetic algorithms perform well. – shadowtalker May 05 '16 at 02:48
  • This might get more answers on one of the computer science or math stack exchange sites (it might be research level, making it appropriate for MathOverflow or theoretical computer science, but don't take my word for it). – Chill2Macht Mar 17 '17 at 22:32

0 Answers0